The Science
From Nobel Prize-winning algorithms to cutting edge econometrics, our foundation is rooted in scientific research.
The Algorithm
Optimally bringing together two sides of a market is a key problem of economics. In the 1960s, David Gale and Lloyd Shapley developed the “Deferred Acceptance” algorithm (DA) to find stable matching among two groups, often called the “stable marriage problem.” In the 1980s, Alvin Roth built on their work and developed systems for matching doctors with hospitals, school children with schools, and organ donors with patients. In the 2010s, Josh Angrist and Parag Pathak built on their work to develop algorithms optimized for assigning students to schools, while accounting priorities and diversity factors, such as sibling preference, gender balance, and diversity. All modern systems use some version of Angrist and Pathak’s work, but only Avela’s is easily configurable, auditable, and equitable.
The Assessment
Random controlled trials (RCT) are the simplest type of experiment to determine the effect of a treatment (such as attending a school) on a population. But we cannot and should not experiment on our kids -their future is too precious to leave to chance. And so, school assignments are not purely random. But, when all the factors that go into the assignment are known, Josh Angrist can use those as a sort of quasi-experiment using instrumental variables (IV) to estimate causal relationships. This approach was first derived by Philip Wright in the 1920s, and Angrist built on that work by developing a two-stage least squares (2SLS) equivalent of the efficient Wald estimator.
Our Team’s Research and Publications
